About Me

Hi, I am Dongyeong Kim. I am a current undergraduate student at KAIST, majoring in the School of Computing and minoring in the Department of Brain and Cognitive Science. I am very interested in the capability of generalisation that might have been shown in autonomous agents and already has been shown in our minds.

Previously, I had a research internship in the Machine Learning and Mind Lab with the supervision of Professor Sungjin Ahn. I have been greatly influenced by him about Model-Based Reinforcement Learning and Object-Centric Learning.

Why Generalisation?

For multiple billion years, at least following the state-of-the-art theory, our universe somehow has been optimising the combination of atoms carefully resulting in higher life forms like us. This process of optimisation not only relies upon the genetic crossovers and random mutations for varying and changing environments but also the generalisation ability of the species, specifically for humans to reduce the time of trying by composing it based on our tools such as logic system which is the result of the Generalisation. Therefore I believe that deepening our understanding of Generalisation is the key to understand our mind and intelligence and which might help to discover our world and its properties more rapidly.

How can Generalisation be connected with current research topics of AI?

There are a bunch of connections in current research, but currently, I am focusing on the following research topics:

• Sample Efficiency in Reinforcement Learning
• Structural Representation Learning
• Zero-shot Learning (connected with search process)

For more info

If you are interested, please contact me via Email or direct message on Bluesky or Twitter (X).